TMA paper Choi P, et al.
1
Examination of Oral Cancer Biomarkers by Tissue Microarray Analysis 1
Peter Choi,1 M.D, Ph.D; C. Diana Jordan,2 M.D.; Eduardo Mendez,1,3 M.D.; John Houck,4 BS; 2
Bevan Yueh,5 M.D.; D. Gregory Farwell,6 M.D.; Neal Futran,1 M.D., D.M.D.; Chu Chen,1,4,7 3
Ph.D. 4
1 Department of Otolaryngology – Head and Neck Surgery, University of Washington, Seattle, 5
WA 98195 6
2 Department of Pathology, University of Washington, Seattle, WA 98195 7
3 Veterans Affairs Puget Sound Health Care System, Seattle, WA 98108 8
4 Program in Epidemiology, Division of Public Health Sciences, Fred Hutchinson Cancer 9
Research Center, Seattle, WA 98109 10
5 Department of Otolaryngology – Head and Neck Surgery, University of Minnesota, 11
Minneapolis, MN 55455 12
6 Department of Otolaryngology – Head and Neck Surgery, University of California - Davis, 13
Sacramento, CA 95820 14
7 Department of Epidemiology, University of Washington, Seattle, WA 98195 15
16 Corresponding Author: Chu Chen, Ph.D. 17 Fred Hutchinson Cancer Research Center 18 1100 Fairview Ave North 19 PO Box 19024 20
M5-C800 21 Seattle, WA 98109 22 Phone (206) 667-6644 23
Fax (206) 667-2537 24 e-mail: [email protected] 25
Word count: 3557 26
27
28
TMA paper Choi P, et al.
2
ABSTRACT 29
Background. Oral squamous cell carcinoma (OSCC) is a major healthcare problem worldwide. 30
Efforts in our laboratory and others focusing on the molecular characterization of OSCC tumors 31
with the use of DNA microarrays have yielded heterogeneous results. To validate the DNA 32
microarray results on a subset of genes from these studies that could potentially serve as 33
biomarkers of OSCC, we elected to examine their expression by an alternate quantitative method 34
and by assessing their protein levels. 35
Design. Based on DNA microarray data from our lab and data reported in the literature, we 36
identified six potential biomarkers of OSCC to investigate further. We employed quantitative, 37
real-time polymerase chain reaction (qRT-PCR) to examine expression changes of CDH11, 38
MMP3, SPARC, POSTN, TNC, TGM3 in OSCC and normal control tissues. We further 39
examined validated markers on the protein level by immunohistochemistry (IHC) analysis of 40
OSCC tissue microarray (TMA) sections. 41
Results. qRT-PCR analysis revealed up-regulation of CDH11, SPARC, POSTN, and TNC gene 42
expression, and decreased TGM3 expression in OSCC compared to normal controls. MMP3 was 43
not found to be differentially expressed. In TMA IHC analyses, SPARC, periostin, and tenascin 44
C exhibited increased protein expression in cancer compared to normal tissues, and their 45
expression was primarily localized within tumor-associated stroma rather than tumor epithelium. 46
Conversely, transglutaminase-3 protein expression was found only within keratinocytes in 47
normal controls, and was significantly down-regulated in cancer cells. 48
Conclusions. Of six potential gene markers of OSCC, initially identified by DNA microarray 49
analyses, differential expression of CDH11, SPARC, POSTN, TNC, and TGM3 were validated by 50
TMA paper Choi P, et al.
3
qRT-PCR. Differential expression and localization of proteins encoded by SPARC, POSTN, 51
TNC, and TGM3 were clearly shown by TMA IHC. 52
53
Introduction 54
Head and neck squamous cell carcinoma (HNSCC) is the 5th most common cancer 55
worldwide.1 The American Cancer Society estimates that approximately 30,990 Americans were 56
diagnosed with and 7,430 died of cancer of the oral cavity and pharynx in 2006.2 Despite 57
considerable advances in the treatment of HNSCC over the past two decades, overall disease 58
outcomes have only modestly improved.2 Local tumor recurrence affects approximately 60% of 59
patients and metastases develops in 15-25%.3 Less than 30% of HNSCC patients experience 60
three or more years of disease-free survival, and many suffer from impaired speech, swallowing, 61
and/or breathing due to the sensitive location of HNSCC tumors within the upper aerodigestive 62
tract.4 Of the various subgroups of HNSCC, oral squamous cell carcinoma (OSCC) is the most 63
common, representing about 75% of all HNSCC cases.2 High throughput investigation into the 64
molecular characteristics of HNSCC has mainly utilized DNA microarray technology to search 65
for gene expression profiles associated with disease and disease outcomes. The literature on 66
DNA microarray profiling of HNSCC shows heterogeneity in the specific genes that were found 67
to be up- or down-regulated in HNSCC. After comparing results from multiple studies, we 68
reported a list of genes commonly found to have dysregulated expression in HNSCC tumors.5 69
Only a handful of these gene expression alterations have been validated by alternate 70
experimental methodologies such as qRT-PCR, Western blot, Northern blot, and IHC. Even 71
fewer have been examined for their correlation with disease severity and metastasis status. 72
TMA paper Choi P, et al.
4
Based on various selection criteria (see Materials & Methods), we selected six genes to 73
analyze further: CDH11, MMP3, SPARC, POSTN, TNC, and TGM3. CDH11 encodes an 74
integral membrane protein, cadherin-11, which mediates cell-cell adhesion and thought to be 75
involved in bone cell differentiation and bone formation.6 MMP3 encodes a secreted protease, 76
matrix metalloproteinase-3, whose action is to degrade the major components of the extracellular 77
matrix (ECM), and is thought to be associated with cervical lymph node metastases in HNSCC.7 78
SPARC (secreted protein, acidic, rich in cysteine), encodes an ECM-associated protein, a.k.a. 79
osteonectin, that inhibits cell-cycle progression, causes changes in cell shape, and influences 80
ECM synthesis.8 SPARC has also been found to be an independent prognostic marker for short 81
disease-free interval and poor overall survival in HNSCC patients.9 POSTN encodes the protein, 82
periostin, which is a ligand for various integrins and as such, supports adhesion and migration of 83
epithelial cells.10 Periostin is thought to promote invasion and angiogenesis in OSCC.11,12 TNC 84
encodes an ECM protein, tenascin-C, that regulates cell adhesion, migration, and growth.13 85
TGM3 encodes transglutaminase-3, which crosslinks intracellular structural proteins and is 86
important in cell envelope formation of the epidermis.14 Transglutaminase-3 is expressed 87
normally in terminally differentiated epithelial cells.15 It has been shown to be down-regulated 88
in esophageal squamous cell carcinoma16-18 and in the progression of oral leukoplakia to 89
OSCC.19 CDH11, MMP3, SPARC, POSTN, and TNC have all been shown in gene microarray 90
experiments to be significantly upregulated in cancer tissues compared to normal controls, 91
whereas TGM3 is significantly down-regulated.5 92
93
94
95
TMA paper Choi P, et al.
5
Materials & Methods 96
Biomarker Selection. The criteria for choosing potential OSCC markers for the current study 97
were: 1) genes from our own DNA microarray data that show the highest Z-scores and greatest 98
expression fold changes between cancer and normal;20 2) these genes have been shown to have 99
significantly altered expression in OSCC when compared with non-cancer tissues in at least four 100
other laboratories, and 3) the genes had available antibodies against their encoded protein 101
products for use in IHC analyses. 102
103
qRT-PCR. Differential gene expression of CDH11, MMP3, SPARC, POSTN, TNC, and TGM3 104
between normal and cancer specimens was quantified by using SYBR® Green I technology and 105
melting-point dissociation curve analyses per manufacturer protocol (Applied Biosystems, Foster 106
City, CA). Total RNA extracted from six normal tissue samples and six tumor tissue samples 107
(Table 1) were used as templates in RT-PCR reactions to generate cDNA. Each sample was 108
divided into five wells for the qRT-PCR reactions: three for the gene of interest and two for the 109
endogenous control, glyceraldehyde 3-phosphate dehydrogenase (GAPDH). qRT-PCR analyses 110
were performed on an ABI 5700 Sequence Detector using 10 ng of cDNA and gene specific 111
primers in 1 × SYBR® Green I PCR Master Mix in a 50-µL reaction. Cycling parameters were 112
50 °C for 2 minutes, 95 °C for 10 minutes, and 40 cycles at 95 °C for 15 seconds and at 60 °C 113
for 1 minute. Primer sequences were designed using PE/ABI Primer Express® software, checked 114
for specificity against the National Center for Biotechnology Information nucleotide data base, 115
and were as follows: CDH11 forward, 5-GCT CAA CCA GCA GAG ACA TTC C-3; CDH11 116
reverse, 5-AGA ATG CAG CTG TCA CCC CTT-3; MMP3 forward, 5-GGC AAG ACA GCA 117
AGG CAT AGA-3; MMP3 reverse, 5-TGG ATA GGC TGA GCA AAC TGC-3; SPARC 118
TMA paper Choi P, et al.
6
forward, 5-CGG CTT TGT GGA CAT CCC TA-3; SPARC reverse, 5-GGA AGG ACT CAT 119
GAC CTG CAT C-3; POSTN forward, 5-ACA ACG CAG CGC TAT TCT GAC-3; POSTN 120
reverse, 5-ATC CAA GTT GTC CCA AGC CTC-3; TNC forward, 5-AGA AAG TCA TCC 121
GGC ACA AGC-3; TNC reverse, 5-ACT CCA GAT CCA CCG AAC ACT G-3; TGM3 forward, 122
5-GAC AAG CGC ATC ACA CAG ACA-3; TGM3 reverse, 5-TCT TTC GTT AGA GCC AAG 123
GCC-3. Melting-curve analyses were run immediately after cycling to verify specificity of the 124
reactions. Quantification of the transcripts was determined by choosing a fluorescence threshold 125
at which the amplification of the target gene was exponential in both tumor samples and normal 126
samples. The PCR cycle number at which the amplification curve intercepted the threshold is 127
termed the threshold cycle (CT). The threshold cycle is inversely proportional to the copy 128
number of the target template. Relative fold changes were calculated by 2- ΔΔCT, where ΔΔCT = 129
[average CT, gene j - average CT, GAPDH] tumor tissue - [average CT, gene j - average CT, 130
GAPDH] normal tissue. 131
132
TMA Construction. TMA blocks (one master and one copy) were constructed with the use of an 133
automated tissue arrayer per manufacturer protocol (Beecher Instruments, Sun Prairie, WI). 134
Formalin-fixed, paraffin-embedded tissue specimens were obtained from the University of 135
Washington, Department of Pathology under Institutional Review Board approval by the Human 136
Subjects Division at the University of Washington and the Fred Hutchinson Cancer Research 137
Center. Our TMA is comprised of 63 tissue specimens from 31 patients (Table 2). Patients 138
ranged from 28 to 88 years (60±15) in age and three quarters were male (74.2%). Tissues came 139
from the oral cavity, oropharynx, and lymph node metastases. Twenty-four of the tissue 140
specimens were from primary OSCC tumors, 16 were from normal tissue taken from cancer 141
TMA paper Choi P, et al.
7
patients, 17 were from cervical lymph node metastases, three were premalignant lesions, and 142
three were from patients without cancer (normal patients). Of the 24 tumors, 4 were stage I/II 143
and 20 were stage III/IV, 12 were T1/T2 and 12 were T3/T4, and 18 had associated cervical 144
lymph node metastases while 6 were non-metastatic. Cores were arrayed in quadruplicate, with 145
a diameter of 0.6 mm. TMA blocks were stored in a nitrogen chamber for antigen preservation. 146
Sections were cut at 5 µM for IHC analyses. 147
148
Immunohistochemistry. Genes that satisfied the following criteria were studied further by IHC 149
analysis of OSCC TMA sections: 1) gene expression in cancer tissues had to be significantly 150
different than in normal tissues by each of the two different methods, qRT-PCR and GeneChip® 151
analysis, and 2) the relative gene expression determined by these two different methods had to be 152
significantly correlated with one another. 153
SPARC IHC was performed using mouse monoclonal anti-SPARC antibody (US 154
Biological, Swampscott, MA) at a dilution of 1:2000, following antigen retrieval consisting of 20 155
minutes of non-pressurized steam incubation in 10 mM citrate buffer, pH 6.0. Periostin IHC was 156
performed using rabbit polyclonal anti-periostin antibody (BioVendor, Candler, NC; 157
http://biovendor.com/pdf/RD181045050.pdf) at a dilution of 1:900. Tenascin-C IHC was 158
performed using mouse monoclonal anti-tenascin-C antibody (BioVendor, Candler, NC; 159
http://biovendor.com/pdf/RE11370C100.pdf) at a dilution of 1:50 following antigen retrieval 160
consisting of two, five-minute microwave incubations in 10 mM citrate buffer, pH 6.0. 161
Transglutaminase-3 IHC was performed using mouse monoclonal antibody (a generous gift from 162
Dr. Kiyotaka Hitomi, Nagoya University, Nagoya, Japan) raised against purified, recombinant 163
human TGase-3,15 at a dilution of 1:1000. TMA sections incubated with antibody diluent alone, 164
TMA paper Choi P, et al.
8
phosphate buffered saline (PBS) containing 1% bovine serum albumin, served as negative 165
controls to confirm specificity of immunostaining for all markers. 166
Slides were deparaffinized by three changes of xylene, seven minutes each, and 167
rehydrated by three changes of 100% ethanol x two minutes, two changes of 95% ethanol x two 168
minutes, and one change of 70% ethanol x one minute. Endogenous peroxidase activity was 169
blocked by incubation in 0.3% H2O2 at room temperature for 10 minutes. Non-specific binding 170
sites were subsequently blocked with 2% normal goat serum (Vector Laboratories, Burlingame, 171
CA) in PBS, pH 7.4, at room temperature for 15 minutes. Slides were then washed and 172
incubated with primary antibody. Washes consisted of a five-minute soak in PBS, pH 7.4, 173
followed by 10 dips in PBS with 1% BSA and 0.01% Triton-X-100 (Sigma, St. Louis, MO), pH 174
7.4, and a second five-minute wash in PBS, pH 7.4. TMA slides were then incubated with 175
primary antibody at the above listed dilutions in PBS with 1% BSA, pH 7.4, at room temperature 176
for one hour in a humidity chamber (Shandon Lipshaw, Inc, Pittsburgh, PA). Sections were then 177
washed again as before, incubated with either biotinylated goat anti-mouse or biotinylated goat 178
anti-rabbit secondary antibody (Vector Laboratories, Burlingame, CA), washed again, and 179
subsequently incubated with ABC Elite reagent (Vector Laboratories, Burlingame, CA) per 180
manufacturer protocol. Slides were then stained by incubation with 0.08% diaminobenzidine, 181
0.01% FeCl3 in 0.05 M Tris buffer, pH 8.0 (Sigma, St. Louis, MO) at 37°C for seven minutes. 182
Sections were counterstained with Mayers Hematoxylin (Dako Cytomation, Carpinteria, CA), 183
dehydrated, mounted with Cytoseal 60 (Richard-Allan Scientific, Kalamazoo, MI) and covered 184
with a coverslip. 185
186
TMA paper Choi P, et al.
9
Scoring of IHC results. Images of the IHC-stained TMA sections were digitized using the 187
BLISS Tracer imaging system and visualized with WebSlide Server software (Bacus 188
Laboratories, Lombard, IL). Marker immunoreactivity was scored using a validated, modified 189
H-score system21 by a board-certified pathologist (C.D.J.), blinded to all characteristics of the 190
cases and controls. IHC scores were determined by taking the product of the estimated staining 191
intensity (0, 1+, 2+, 3+) and area of tissue (tumor or normal; epithelial or stromal) stained (0% = 192
0, <25% = 1, 25-75% = 2, >75% = 3), giving a range of possible scores between 0 and 9. IHC 193
scores for replicate cores were averaged to determine a composite score for each case. 194
195
Statistical Analysis. qRT-PCR results were analyzed by an unpaired Student’s t-test to compare 196
gene expression between cancer and normal specimens. Relative gene expression values 197
determined by qRT-PCR analyses were then compared with those previously determined by gene 198
microarray analyses20 by linear regression analysis. Pairwise comparisons of IHC scores were 199
made between primary tumor tissue and normal mucosa, tumor tissue and lymph nodes, tumor 200
tissue and CIS, and CIS tissue and normal mucosa, using unpaired Student’s t-tests. Because of 201
multiple comparisons, we felt that a threshold of P<.05 was too low, and considered comparisons 202
to be statistically significant only if P<.01. All statistical analyses were performed using Stata 203
9.0 software (StataCorp, College Station, TX). 204
205
Results 206
CDH11, SPARC, POSTN, TNC, and TGM3 exhibited significant differences in 207
expression between cancer and normal specimens by qRT-PCR (Table 3). There was a trend 208
TMA paper Choi P, et al.
10
towards up-regulation of MMP3 expression in cancer compared to normal specimens, but this 209
did not reach statistical significance. 210
Results of the linear regression analyses show good correlation between relative gene 211
expression as determined by qRT-PCR and that previously determined by DNA microarray20 on 212
the same specimens for MMP3, SPARC, POSTN, TNC, and TGM3 (Table 3). Correlation 213
between qRT-PCR and gene microarray expression data for CDH11 did not reach statistical 214
significance (r = 0.45, p ≤ 0.14). 215
Representative TMA cores stained for SPARC, periostin, tenascin-C, and 216
transglutaminase-3 are shown for both normal mucosa and primary OSCC tumors (Figure 1). 217
Staining with antibodies to SPARC was predominantly localized to vessels, fibroblasts, and 218
subsets of carcinoma cells. Virtually no epithelial cell staining was observed in normal mucosa 219
(Figure 1A). SPARC IHC staining in epithelium, fibroblasts, and vessels was significantly 220
higher in tumor specimens compared to normal controls (Figure 1B, Table 4). There were no 221
significant differences with regard to epithelial SPARC expression among tumors of different 222
TNM stage. 223
Staining with antibodies to periostin was predominantly associated with fibroblasts and 224
the ECM. Expression was significantly higher in cancer versus normal controls (Figure 1C-D, 225
Table 4). Epithelial cancer cells in over half (14/24, 58.3%) of the primary tumors and 226
approximately one fourth (4/17, 23.5%) of the cervical metastatic tumors exhibited faint staining, 227
whereas the remaining tumors were completely negative for epithelial cell staining. Of the 228
tumors that were positive for epithelial staining, 13/14 (92.8%) were stage III/IV tumors 229
representing 65% of the 20 stage III/IV tumors examined, whereas only one (7.2%) was an early 230
stage tumor (Case 16, a recurrent T2N0M0), representing one of the four stage I/II tumors 231
TMA paper Choi P, et al.
11
examined. In addition, all 8 (100%) of the T4 tumors on our TMA were positive for epithelial 232
periostin staining. Non-neoplastic epithelium was virtually negative for staining. Staining of 233
carcinoma cells was, on average, higher in primary tumors compared to cervical lymph node 234
metastases (Table 4). Average ECM staining of periostin also was greater in primary tumors 235
compared to that within the metastatic lymph nodes (Table 4). 236
Staining with antibodies to tenascin-C was predominantly localized to the ECM (Figure 237
1E-F). In normal mucosa, only the region of ECM immediately adjacent to the basal epithelium 238
showed moderate staining. Tenascin-C staining in the ECM was significantly higher in OSCC 239
tumors compared to normal mucosa (Table 4). Staining was occasionally associated with 240
carcinoma cells, particularly in tumor regions directly adjacent to desmoplastic stroma. 241
Staining with antibodies to transglutaminase-3 localized only within epithelial cells; no 242
stromal expression was observed (Figure 1G-H). Within non-neoplastic epithelium, the 243
suprabasal layers stained intensely, while the basal layer of epithelial cells was negative for 244
staining. Staining in invasive carcinomas was patchy, and when present, was typically 245
associated with areas of increased differentiation and keratinization. Many carcinomas were 246
completely negative for staining. There were statistically significant differences in 247
transglutaminase-3 immunoreactivity between different specimen groups, such that the highest 248
expression was seen in non-neoplastic epithelium, with significantly lower expression in CIS 249
specimens, and stepwise significantly lower expression in primary tumors (Table 4). 250
251
Discussion 252
We selected six potential biomarkers of OSCC for the current validation study by 253
examining DNA microarray data both from our laboratory, as well as that published in the 254
TMA paper Choi P, et al.
12
literature.5 SPARC, POSTN, TNC, and TGM3 microarray expression differences were validated 255
by both qRT-PCR and IHC of TMA sections. The qRT-PCR results for MMP3 and CDH11 did 256
not reach statistical significance. In contrast with another IHC study reporting high levels of 257
periostin expression within oral carcinoma epithelium,12 we noted periostin staining to be 258
localized primarily to the stroma and did not see robust staining within tumor cells themselves 259
(Figure 1D). The reason for this disparity in IHC staining is unclear. Non-overlapping antibody 260
epitopes may partially explain the disparity of periostin IHC staining patterns. The antibody we 261
used for periostin IHC was raised in rabbits against recombinant human periostin containing 648 262
amino acid residues (corresponding to amino acids 22-669 of full-length periostin) with an N-263
terminal HisTag fusion (http://biovendor.com/pdf/RD181045050.pdf), whereas periostin IHC 264
experiments by Siriwarden et al. utilized a polyclonal antibody generated by immunizing rabbits 265
with a specific peptide (EGEPEFRLIKEGETC) corresponding to amino acids 679-692 of full-266
length periostin.12 267
Despite the strong stromal predominance of periostin expression we observed, the 268
percentage of OSCC tumors positive for epithelial periostin in our study (58%) was comparable 269
to that reported by Siriwarenda et al.12 (69%). A majority (65%) of the stage III/IV OSCC 270
tumors, including 100% of T4 tumors we examined, were positive for epithelial periostin 271
immunostaining, compared to only 25% of the stage I/II tumors. These findings suggest that 272
epithelial expression of periostin may be associated with a more aggressive tumor phenotype in 273
OSCC. This is supported by other studies, which show that subsets of HNSCC cells expressing 274
periostin, or cells engineered to overexpress periostin, exhibit enhanced tumor growth and 275
invasiveness, and tumors that express periostin have a more invasive phenotype.11,12 276
TMA paper Choi P, et al.
13
We found the proportion of metastatic lymph node tumors positive for epithelial periostin 277
expression (23.5%) was less than half that of primary tumors (58.3%). However, each of the 278
positively stained lymph node tumors was associated with a primary tumor that also had 279
epithelial periostin expression, suggesting that presence of periostin in the epithelium of primary 280
tumors may be necessary, but not sufficient, for its presence in metastatic tumors. 281
In oral and laryngeal squamous cell carcinoma, increased levels of tenascin-C 282
immunostaining have been found to correlate with malignancy and invasion22-25 Abundant 283
expression of tenascin-C in our OSCC TMA sections was localized primarily to the stroma, 284
although some minor staining of tumor cells was also observed, particularly at tumor edges 285
adjacent to desmoplastic stroma. This observation is consistent with reports in the literature 286
showing tenascin-C localization along the invasive fronts of carcinomas of the lung, liver, 287
bladder, and skin.26,27 288
Roepman, et al.28 and Schmalbach, et al.29 identified TGM3 to be significantly down-289
regulated in metastatic HNSCC compared to both non-metastatic tumors and normal epithelium. 290
O’Donnell et al.30 similarly found significant down-regulation of TGM3 gene expression in 291
metastatic primary OSCC tumors compared to non-metastatic primaries. Our cross-sectional 292
IHC data show that the levels of transglutaminase-3 protein expression were seen to decrease in 293
a stepwise fashion from normal to premalignant to malignant specimens. This suggests that the 294
loss of transglutaminase-3 activity might be associated with the progression of squamous cell 295
carcinoma. 296
All of the up-regulated gene markers we identified by reviewing gene microarray 297
reports, validated by qRT-PCR, and subsequently studied with IHC revealed protein expression 298
to be localized primarily within the stroma, and modestly or not at all within tumor cells. This 299
TMA paper Choi P, et al.
14
finding illustrates an important point regarding the interpretation of gene microarray data based 300
on the methodology used for specimen processing. The methods employed by different 301
laboratories for tumor specimen processing vary significantly.5 While some investigators 302
isolated relatively homogeneous populations of tumor cells for microarray analysis via laser 303
capture microdissection (LCM), others established arbitrary thresholds for minimum tumor cell 304
content in surgical specimens, as assessed by histologic evaluation of adjacent tissue, prior to 305
RNA extraction and microarray analysis. The latter method clearly results in varying amounts of 306
stromal cells contributing to the final pool of extracted RNA, and thus the variability of the 307
resultant microarray data. Notably, even LCM does not ensure isolation of a purely 308
homogeneous population of tumor cells, as varying degrees of leukocytosis and 309
neovascularization within tumors exist and correlate with survival, tumor stage, metastases, and 310
presence of extracapsular spread in HNSCC.31-33 311
Up-regulation of SPARC, POSTN, or TNC was not reported by any of the DNA 312
microarray studies that examined expression of HNSCC cell lines34-36 or by others that employed 313
LCM to isolate tumor cells from stroma.37-40 Presumably this is due to the relative absence of 314
stromal cells within the analyzed specimens in these studies, although absence of one or more of 315
these markers on the microarrays used by these studies may also contribute. These findings, 316
together with the IHC data we report here, suggest that up-regulation of SPARC, POSTN, and 317
TNC is due to 1) up-regulation within stromal cell populations vs carcinoma cells and/or 2) 318
stroma-induced transcriptional upregulation of these markers in cancer cells. In any case, these 319
observations underscore the importance in examining both tumor and stroma in the pathogenesis 320
of OSCC. 321
TMA paper Choi P, et al.
15
The “seed and soil” hypothesis of tumor-stromal interaction was originally proposed by 322
Paget in 1889, but only recently have researchers examined how tumor microenvironments 323
influence the growth and spread of cancers. Carcinoma-associated fibroblasts (CAFs), ECM 324
macromolecules, neovascularization, and inflammatory and immune cell infiltration within the 325
stroma adjacent to tumors can have profound effects on tumor progression in breast, prostate, 326
and skin carcinomas.41 The situation in OSCC is less well understood, but studies of CAFs, 327
ECM turnover and tumor cell motility have begun to delineate the role of desmoplastic stroma in 328
OSCC carcinogenesis.42-44 Recently, Weber et al.45 performed genome-wide analysis of loss of 329
heterozygosity (LOH) and allelic imbalance (AI) on LCM-isolated specimens of tumor stroma 330
and tumor epithelium from over 120 OSCC patients with a history of smoking. They discovered 331
over 40 hot spots of LOH/AI within the stroma, nearly twice as many as they found in the 332
epithelium, and subsequently identified three stroma-specific loci that were significantly 333
associated with tumor size and cervical lymph node metastasis.45 These findings again highlight 334
the importance of examining both stromal as well as epithelial elements in OSCC, and suggest 335
that stromal alterations play a crucial part in facilitating OSCC invasion and metastasis. 336
337
Conclusions 338
Our observations indicate that significant changes in the expression of four genes, 339
SPARC, POSTN, TNC, and TGM3, initially identified by gene microarray studies, are associated 340
with similar changes in protein expression based on IHC analyses. The localization of SPARC, 341
periostin, tenascin-C predominately within the stroma of OSCC tumors supports the idea that 342
stromal elements are important in OSCC pathogenesis. 343
344
TMA paper Choi P, et al.
16
Acknowledgments. This review was supported by grant R01 CA 095419-01A1 from the 345
National Institutes of Health, National Cancer Institute, Bethesda, MD; National Research 346
Service Award T32 DC00018 from the National Institutes of Health, National Institute on 347
Deafness and Other Communication Disorders, Bethesda, MD; and in part by Fred Hutchinson 348
Cancer Research Center Funds. 349
350
References 351 352
1. Sankaranarayanan,R, Masuyer,E, Swaminathan,R, Ferlay,J, Whelan,S. Head and neck 353 cancer: a global perspective on epidemiology and prognosis. [Review] [14 refs]. Anticancer 354 Res. 1998;18:4779-4786. 355
2. Ries, L. A. G., Harkins, D., Krapcho, M., Mariotto, A., Miller, B. A., Feuer, E. J., Clegg, 356 L., Eisner, M. P., Horner, M. J., Howlader, N., Hankey, B. F., and Edwards, B. K. SEER 357 Cancer Statistics Review 1975-2003. 2006. National Cancer Institute. 358
3. Genden,EM, Ferlito,A, Bradley,PJ, Rinaldo,A, Scully,C. Neck disease and distant 359 metastases. Oral Oncology. 2003;39:207-212. 360
4. Dimery,IW, Hong,WK. Overview of combined modality therapies for head and neck 361 cancer. [Review] [231 refs]. J.Natl.Cancer Inst. 1993;85:95-111. 362
5. Choi,P, Chen,C. Genetic Expression Profiles and Biologic Pathway Alterations in Head 363 and Neck Squamous Cell Carcinoma. Cancer. 2005;104:1113-1128. 364
6. Okazaki,M, Takeshita,S, Kawai,S et al. Molecular cloning and characterization of OB-365 cadherin, a new member of cadherin family expressed in osteoblasts. J.Biol.Chem. 366 1994;269:12092-12098. 367
7. Wiegand,S, Dunne,AA, Muller,HH et al. Metaanalysis of the significance of matrix 368 metalloproteinases for lymph node disease in patients with head and neck squamous cell 369 carcinoma. Cancer. 2005;104:94-100. 370
8. Bradshaw,AD, Graves,DC, Motamed,K, Sage,EH. SPARC-null mice exhibit increased 371 adiposity without significant differences in overall body weight. Proc.Natl.Acad.Sci.U.S.A. 372 2003;100:6045-6050. 373
9. Chin,D, Boyle,GM, Williams,RM et al. Novel markers for poor prognosis in head and neck 374 cancer. Int.J.Cancer. 2005;113:789-797. 375
TMA paper Choi P, et al.
17
10. Gillan,L, Matei,D, Fishman,DA, Gerbin,CS, Karlan,BY, Chang,DD. Periostin secreted by 376 epithelial ovarian carcinoma is a ligand for alpha(V)beta(3) and alpha(V)beta(5) integrins 377 and promotes cell motility. Cancer Res. 2002;62:5358-5364. 378
11. Kudo,Y, Ogawa,I, Kitajima,S et al. Periostin promotes invasion and anchorage-379 independent growth in the metastatic process of head and neck cancer. Cancer Res. 380 2006;66:6928-6935. 381
12. Siriwardena,BS, Kudo,Y, Ogawa,I et al. Periostin is frequently overexpressed and 382 enhances invasion and angiogenesis in oral cancer. Br.J.Cancer. 2006;95:1396-1403. 383
13. Jones,FS, Jones,PL. The tenascin family of ECM glycoproteins: structure, function, and 384 regulation during embryonic development and tissue remodeling. Dev.Dyn. 2000;218:235-385 259. 386
14. Kim,IG, Gorman,JJ, Park,SC, Chung,SI, Steinert,PM. The deduced sequence of the novel 387 protransglutaminase E (TGase3) of human and mouse. J.Biol.Chem. 1993;268:12682-388 12691. 389
15. Hitomi,K, Presland,RB, Nakayama,T, Fleckman,P, Dale,BA, Maki,M. Analysis of 390 epidermal-type transglutaminase (transglutaminase 3) in human stratified epithelia and 391 cultured keratinocytes using monoclonal antibodies. J.Dermatol.Sci. 2003;32:95-103. 392
16. Luo,A, Kong,J, Hu,G et al. Discovery of Ca2+-relevant and differentiation-associated 393 genes downregulated in esophageal squamous cell carcinoma using cDNA microarray. 394 Oncogene. 2004;23:1291-1219. 395
17. Liu,W, Yu,ZC, Cao,WF, Ding,F, Liu,ZH. Functional studies of a novel oncogene TGM3 in 396 human esophageal squamous cell carcinoma. World J of Gastroenterol. 2006;12:3929-397 3932. 398
18. Chen,BS, Wang,MR, Xu,X et al. Transglutaminase-3, an esophageal cancer-related gene. 399 Int.J.Cancer. 2000;88:862-865. 400
19. Ohkura,S, Kondoh,N, Hada,A et al. Differential expression of the keratin-4, -13, -14, -17 401 and transglutaminase 3 genes during the development of oral squamous cell carcinoma 402 from leukoplakia. Oral Oncol. 2005;41:607-613. 403
20. Mendez,E, Cheng,C, Farwell,DG et al. Transcriptional expression profiles of oral 404 squamous cell carcinomas. Cancer. 2002;95:1482-1494. 405
21. Budwit-Novotny,DA, McCarty,KS, Cox,EB et al. Immunohistochemical analyses of 406 estrogen receptor in endometrial adenocarcinoma using a monoclonal antibody. Cancer 407 Res. 1986;46:5419-5425. 408
22. Zidar,N, Gale,N, Kambic,V, Fischinger,J. Expression of tenascin and fibronectin in benign 409 epithelial hyperplastic lesions and squamous carcinoma of the larynx. Anticancer Res. 410 2001;21:451-454. 411
TMA paper Choi P, et al.
18
23. Juhasz,A, Bardos,H, Repassy,G, Adany,R. Characteristic distribution patterns of tenascin 412 in laryngeal and hypopharyngeal cancers. Laryngoscope. 2000;110:84-92. 413
24. Hindermann,W, Berndt,A, Borsi,L et al. Synthesis and protein distribution of the unspliced 414 large tenascin-C isoform in oral squamous cell carcinoma. J.Pathol. 1999;189:475-480. 415
25. Regezi,JA, Ramos,DM, Pytela,R, Dekker,NP, Jordan,RC. Tenascin and beta 6 integrin are 416 overexpressed in floor of mouth in situ carcinomas and invasive squamous cell carcinomas. 417 Oral Oncol. 2002;38:332-336. 418
26. Dang,C, Gottschling,M, Roewert,J, Forschner,T, Stockfleth,E, Nindl,I. Tenascin-C patterns 419 and splice variants in actinic keratosis and cutaneous squamous cell carcinoma. 420 Br.J.Dermatol. 2006;155:763-770. 421
27. Chiquet-Ehrismann,R, Chiquet,M. Tenascins: regulation and putative functions during 422 pathological stress. J.Pathol. 2003;200:488-499. 423
28. Roepman,P, Wessels,LF, Kettelarij,N et al. An expression profile for diagnosis of lymph 424 node metastases from primary head and neck squamous cell carcinomas. Nat.Genet. 425 2005;37:182-186. 426
29. Schmalbach,CE, Chepeha,DB, Giordano,TJ et al. Molecular profiling and the identification 427 of genes associated with metastatic oral cavity/pharynx squamous cell carcinoma. 428 Arch.Otolaryngol.Head Neck Surg. 2004;130:295-302. 429
30. O'donnell,RK, Kupferman,M, Wei,SJ et al. Gene expression signature predicts lymphatic 430 metastasis in squamous cell carcinoma of the oral cavity. Oncogene. 2005;24:1244-1251. 431
31. Marcus,B, Arenberg,D, Lee,J et al. Prognostic factors in oral cavity and oropharyngeal 432 squamous cell carcinoma. Cancer. 2004;101:2779-2787. 433
32. Reichert,TE, Scheuer,C, Day,R, Wagner,W, Whiteside,TL. The number of intratumoral 434 dendritic cells and zeta-chain expression in T cells as prognostic and survival biomarkers in 435 patients with oral carcinoma. Cancer. 2001;91:2136-2147. 436
33. Badoual,C, Hans,S, Rodriguez,J et al. Prognostic value of tumor-infiltrating CD4+ T-cell 437 subpopulations in head and neck cancers. Clin Cancer Res. 2006;12:465-472. 438
34. Pawelec,G. Tumour escape: antitumour effectors too much of a good thing?. [Review] [172 439 refs]. Cancer Immunol.Immunother. 2004;53:262-274. 440
35. Jeon,GA, Lee,JS, Patel,V et al. Global gene expression profiles of human head and neck 441 squamous carcinoma cell lines. Int.J.Cancer. 2004;112:249-258. 442
36. Moriya,T, Seki,N, Shimada,K et al. In-house cDNA microarray analysis of gene expression 443 profiles involved in SCC cell lines. Int. J. of Mol. Med. 2003;12:429-435. 444
TMA paper Choi P, et al.
19
37. Irie,T, Aida,T, Tachikawa,T. Gene expression profiling of oral squamous cell carcinoma 445 using laser microdissection and cDNA microarray. Med. Electron Microscopy. 2004;37:89-446 96. 447
38. Leethanakul,C, Patel,V, Gillespie,J et al. Distinct pattern of expression of differentiation 448 and growth-related genes in squamous cell carcinomas of the head and neck revealed by 449 the use of laser capture microdissection and cDNA arrays. Oncogene. 2000;3220-3224. 450
39. Leethanakul,C, Knezevic,V, Patel,V et al. Gene discovery in oral squamous cell carcinoma 451 through the Head and Neck Cancer Genome Anatomy Project: confirmation by microarray 452 analysis. Oral Oncol. 2003;39:248-258. 453
40. Hwang,D, Alevizos,I, Schmitt,WA et al. Genomic dissection for characterization of 454 cancerous oral epithelium tissues using transcription profiling. Oral Oncol.. 2003;39:259-455 268. 456
41. Mueller,MM, Fusenig,NE. Friends or foes – bipolar effects of the tumour stroma in cancer. 457 Nat Rev Cancer 2004;4:839-849. 458
42. Rosenthal,EL, McCrory,A, Talbert,M, Carroll,W, Magnuson,JS, Peters,GE. Expression of 459 proteolytic enzymes in head and neck cancer-associated fibroblasts. 460 Arch.Otolaryngol.Head Neck Surg. 2004;130:943-947. 461
43. Rosenthal,E, McCrory,A, Talbert,M, Young,G, Murphy-Ullrich,J, Gladson,C. Elevated 462 expression of TGF-beta1 in head and neck cancer-associated fibroblasts. Mol.Carcinog. 463 2004;40:116-121. 464
44. Ziober,AF, Falls,EM, Ziober,BL. The extracellular matrix in oral squamous cell carcinoma: 465 friend or foe? Head & Neck. 2006;28:740-749. 466
45. Weber,F, Xu,Y, Zhang,L et al. Microenvironmental genomic alterations and 467 clinicopathological behavior in head and neck squamous cell carcinoma. JAMA. 468 2007;297:187-195. 469
470
471
472
473
474
475
476
477
TMA paper Choi P, et al.
20
Figure 1 – IHC of OSCC TMA sections 478
Representative tissue cores are shown. (A) SPARC IHC of normal mucosa – Occasional faint 479
staining of vessels and fibroblasts was identified. Virtually no epithelial staining was seen. (B) 480
SPARC IHC of a T4N2bM0 primary OSCC – Increased staining was observed, mostly within 481
vessels and fibroblasts of the stroma, but also within some carcinoma cells. (C) Periostin IHC 482
of normal mucosa – Focal faint-to-moderate vascular staining and very focal faint subepithelial 483
stromal staining was identified. (D) Periostin IHC of T4N2bM0 primary OSCC – Strong 484
staining localized primarily within the ECM and stromal fibroblasts. Many late stage tumors 485
also exhibited modest staining within cancer epithelium, as in this case. (E) Tenascin-C IHC of 486
normal mucosa – A narrow zone of faint-to-moderate stromal staining was identified beneath the 487
basal layer of normal epithelium. (F) Tenascin-C IHC of T4N2bM0 primary OSCC – Strong 488
staining was observed mostly within the carcinoma-associated stroma. (G) Transglutaminase-3 489
IHC of normal mucosa – Moderate-to-strong staining was seen in suprabasal epithelium of all 490
normal mucosa. (H) Transglutaminase-3 IHC of T4N2bM0 primary OSCC – Although some 491
OSCC tumors exhibited faint epithelial staining, primarily in more differentiated and keratinized 492
regions, many tumors did not stain at all, as in this case. 493
494
495
496
497
498
499
500
TMA paper Choi P, et al.
21
501
Table 1 – Characteristics of Specimens Used in qRT-PCR Analyses 502 503
Specimen Age Sex Diagnosis Stage Grade
Normal 43 M Obstructive sleep apnea T0N0M0 N/A Normal 37 M Obstructive sleep apnea T0N0M0 N/A Normal 48 F Obstructive sleep apnea T0N0M0 N/A
Normal 48 M Buccal mucosa from patient with lateral tongue SCCA T2N0M0 N/A
Normal 54 F Buccal mucosa from patient with lateral tongue SCCA T2N0M0 N/A
Normal 64 M Palate mucosa from patient with tonsil/palate SCCA T2N2bM0 N/A
Cancer 48 M Lateral tongue SCCA T2N0M0 moderate Cancer 81 F Retromolar trigone SCCA T4N2bM0 moderate-to-poor Cancer 63 M Tonsil/soft palate SCCA T2N2bM0 moderate Cancer 61 M Retromolar trigone SCCA T4N2bM0 moderate Cancer 76 M Gingivobuccal sulcus SCCA T4N0M0 moderate Cancer 54 F Lateral tongue SCCA T2N0M0 moderate
SCCA = squamous cell carcinoma 504 505 506
507
508
509
510
511
512
513
514
515
516
517
TMA paper Choi P, et al.
22
Table 2 – Characteristics of Tissue Microarray Specimens 518 519
Case Specimen(s) Age Sex Diagnosis Stage Grade 1 NLn 43 M Obstructive sleep apnea T0N0M0 N/A
2 NLn 37 M Obstructive sleep apnea T0N0M0 N/A
3 NLn 48 F Obstructive sleep apnea T0N0M0 N/A
4 T, LN 51 F Lateral tongue SCCA T4N2bM0 well-to-moderate
5 NLc, T 48 M Lateral tongue SCCA T2N0M0 moderate
6 T, LN 81 F Retromolar trigone SCCA T4N2bM0 moderate-to-poor
7 NLc, T, LN 46 F Lateral tongue SCCA T2N2bM1 well
8 NLc, T, LN 63 M Tonsil/soft palate SCCA T2N2bM0 moderate
9 T, LN 61 M Retromolar trigone SCCA T4N2bM0 moderate
10 NLc, T 76 M Gingivobuccal sulcus SCCA T4N0M0 moderate
11 NLc, T, LN 54 M Tonsil/soft palate SCCA T4N2bM0 moderate
12 NLc, T, LN 77 M Floor of mouth SCCA T1N2bM0 well-to-moderate
13 NLc, T 54 F Lateral tongue SCCA T2N0M0 moderate
14 CIS, NLc 28 M Lateral tongue leukoplakia TisN0M0 CIS
15 CIS 83 F Lower lip leukoplakia TisN0M0 CIS
16 NLc, T 39 F Recurrent floor of mouth SCCA T2N0M0 well
17 T 51 M Floor of mouth SCCA T2N0M0 moderate
18 CIS 49 M Floor of mouth leukoplakia TisN0M0 CIS
19 NLc, T 50 M Lateral tongue SCCA T2N0M0 moderate
20 LN 57 M Posterior tongue SCCA T1N2bM0 not available
21 T, LN 88 F Retromolar trigone SCCA T4N2bM0 well
22 NLc, T, LN 61 M Base of tongue SCCA T3N2bM0 well-to-moderate
23 NLc, T 63 M Lateral tongue SCCA T1N2bM0 moderate-to-poor
24 T, LN 62 M Recurrent lateral tongue SCCA T2N2bM0 poor
25 NLc, T, LN 52 M Anterior tongue SCCA T3N1M0 well
26 NLc, T, LN 62 M Lateral tongue SCCA T2N2bM0 moderate
27 NLc, T, LN 62 M Base of tongue SCCA T3N2bM0 moderate
28 T, LN 70 M Base of tongue SCCA T3N2bM0 poor
29 T, LN 76 M Lateral tongue SCCA T4N2bM0 moderate
30 NLc, T, LN 85 M Lateral tongue SCCA T3N2bM0 moderate
31 T 81 M Lateral tongue SCCA T4N0M0 well
NLc = matched normal from cancer/CIS patients, NLn = unmatched normal from normal patients, 520 T = primary tumor, LN = lymph node metastasis, SCCA = squamous cell carcinoma 521
TMA paper Choi P, et al.
23
Table 3 - qRT-PCR Validation of Gene Microarray Data 522 523 Gene Method CA:NML Ratio1 Linear Correlation CDH11 microarray 19.8 (p=0.003) r = 0.45, p ≤ 0.14 qRT-PCR 18.9 (p=0.03) MMP3 microarray 22.9 (p=0.09) r = 0.97, p ≤ 0.0004 qRT-PCR 4.30 (p=0.10) SPARC microarray 5.67 (p=0.001) r = 0.91, p ≤ 0.0002 qRT-PCR 3.90 (p=0.001) POSTN microarray 18.8 (p=0.008) r = 0.90, p ≤ 0.0003 qRT-PCR 4.52 (p=0.009) TNC microarray 7.80 (p=0.0003) r = 0.90, p ≤ 0.0003 qRT-PCR 6.67 (p<0.0001) TGM3 microarray 0.057 (p<0.0001) r = 0.63, p ≤ 0.03 qRT-PCR 0.068 (p=0.03) 1CA:NML = cancer-to-normal 524 525
526
527
528
529
530
531
532
533
534
535
536
TMA paper Choi P, et al.
24
Table 4 – Analysis of IHC Scores for SPARC, Periostin, Tenascin-C, and 537 Transglutaminase-3 538 539
540
Columns A-D list IHC scores for the specified tissues. P-values for unpaired student t-tests are 541 listed in the remaining columns. 542 CIS = carcinoma in-situ, ECM = extracellular matrix, LN = lymph node, TGM3 = 543 transglutaminase-3, TNC = tenascin-C 544
A Primary Tumor
B
Normal
C LN
Tumor
D
CIS A vs B P-value
A vs C P-value
A vs D P-value
B vs D P-value
TNCepithelium 1.45 0.07 1.13 1.00 .0009 .5720 .5215 .5215 TNCECM 5.09 2.61 4.01 4.00 .0084 .3590 .0333 .0333 Periostinepithelium 1.28 0.45 0.38 0.83 .1057 .0181 .6214 .6214 PeriostinECM 5.76 3.38 3.34 4.00 .0017 .0037 .8087 .8087 Periostinfibroblast 5.48 3.31 3.84 3.25 .0034 .0373 .9714 .9714 SPARCepithelium 1.90 0.27 1.39 1.67 .0002 .4673 .3005 .3005 SPARCfibroblast 8.56 4.69 6.63 7.00 .0000 .0268 .0312 .0312 SPARCvessels 6.96 4.12 6.70 8.25 .0000 .7081 .0532 .0532 TGM3epithelium 0.31 5.44 0.44 0.92 .0000 .6665 .0000 .0000
Top Related